SBIR Phase II: Intelligent Interactive Guidance System

Award ID: 2604996

The broader impact / commercial potential of this Small Business Innovation Research (SBIR) Phase II project is to improve how claims are analyzed and resolved by turning case file data into clear, verifiable guidance for claims professionals. The innovation will enhance scientific and technological understanding by advancing practical methods for organizing large sets of unstructured documents, linking guidance to source material, and improving confidence in high-stakes decisions. Commercialization is expected through enterprise software-as-a-service deployment and long-term service agreements, creating a durable advantage through explainable analysis, domain-tailored workflows, and scalable performance. The innovation is likely to be a key factor in commercial success because it will reduce review time and could lower operating costs. This Small Business Innovation Research (SBIR) Phase II project advances an intelligent interactive guidance system that analyzes large collections of materials. The problem is that case-critical facts, events, and relationships are distributed across many different unstructured documents, including investigation reports, policies, medical records, exhibits, testimony, and visual evidence, making accurate and comprehensive review slow and tedious. The research objectives are to improve extraction of entities and events, organize evidence into a knowledge graph, generate faithful summaries, answer complex multi-step questions, and expand analysis to photographic and multimedia evidence. The proposed research will combine natural language processing, retrieval-augmented generation, graph-based reasoning, explainable answer generation, timeline construction, user feedback loops, and scalable parallel processing. Anticipated technical results include high retrieval precision, better summarization fidelity, improved source-grounded reasoning over complex case records, strong transparency through links to document locations, and a market-ready system that performs reliably in real-world decision environments. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Program Director: Alastair Monk